地球信息科学学报 ›› 2013, Vol. 15 ›› Issue (6): 862-870.doi: 10.3724/SP.J.1047.2013.00862

所属专题: 地理大数据

• 地理模型与算法 • 上一篇    下一篇

面向大数据的地理格网分析操作模型比较

余劲松弟, 吴升   

  1. 福州大学福建省空间信息工程研究中心, 空间数据挖掘与信息共享教育部重点实验室, 福州 350002
  • 收稿日期:2013-11-02 修回日期:2013-12-07 出版日期:2013-12-25 发布日期:2013-12-25
  • 通讯作者: 吴升(1972-),男,博士,教授,博导,主要从事GIS的开发与应用,数字区域与数字城市的理论、方法和应用研究。E-mail:ws0110@163.com E-mail:ws0110@163.com
  • 作者简介:余劲松弟(1984-),男,博士,副教授,研究方向为地理信息标准化、数字化对地观测、空间数据库和长期科学数据保存。E-mail:yyx350@126.com
  • 基金资助:

    国家“863”计划重大项目课题(2012AA12A208);国家科技支撑计划项目(2007BAH16B00);欧盟FP7项目(283610)。

Research Progress of Array Analytics towards Big Data

YU Jinsongdi, WU Sheng   

  1. Spatial Information Research Center, Fujian Province, Key Laboratory of Spatial Data Mining and Information Sharing, Ministry of Education, Fuzhou University, Fuzhou 350002, China
  • Received:2013-11-02 Revised:2013-12-07 Online:2013-12-25 Published:2013-12-25

摘要:

格网分析操作建模,包括地理格网索引域、值域及其之间的函数定义;基于该定义的操作函数、输入与输出的形式化表达;界定了相应的格网分析技术及其应用。随着物联网和空间对地观测的发展,多维大数据的涌现,对现有2维格网分析基础设施的发展提出了挑战。本文对业界主流的各种格网操作模型进行了系统回顾和比较:首先,介绍地球科学各个领域的多维地理格网及分析需求;然后,抽象了多维格网的形式化定义元素;在此基础上,比较了国内外主要多维格网操作模型,讨论了相应格网操作的输入与输出特征,并总结了不同操作模型的表达能力,论述了格网分析操作建模的方法和需要注意的问题;最后,总结了国内外主流的格网分析语言和系统实现,并展望了其发展趋势和研究方向。

关键词: 信息应用基础设施, 格网分析, 大数据, 地理格网

Abstract:

The development of Internet of Things and Earth Observation brings big observation array data and big array analytics. This paradigm shift challenges the development of current GIS infrastructure which needs innovative proof and design. Array Analytics is considered as one of potential solutions to address today's data tsunami challenge. Formal definitions, such as array domain, range, the mapping function between domain and range, operators, the corresponding input and output are basic building blocks of Array Analytics theory. This paper focuses on the world-wide research progress of Array Analytics towards Big Data. Main Array Analytics models are reviewed and discussed by considering different array definition, operation functions, input and output, making it possible to explore character differentiation and summarize model expressiveness. Then, main array analytics technologies and implementations are discussed and compared by considering different models. Accordingly, array analytics modeling patterns, principle issues and directions for future research are concluded as guidance for further exploration on array analytics theory research, technology improvement and domain applications. National Digital Information Infrastructure Initiatives need effective array storage and analytics platforms. The research is conducted as input of further enhancement on Information Application Infrastructure, which is key backbone of GIS analytics services in terms of earth science system monitoring and forecasting, existing digital construction consolidation, and GIS infrastructure improvement, so as to provide effective support to digital region construction.

Key words: big data, array analytics, information application infrastructure